# Migration from Other Providers
Guide for migrating to xAI from OpenAI, Anthropic, and other providers.
## Overview
The xAI API was designed to be compatible with both the OpenAI and Anthropic SDKs, making migration straightforward.
**Recommended**: Use the native xAI Python SDK or Vercel AI SDK for JavaScript for the best experience and access to all features.
## Migrating from OpenAI
### Minimal Changes Required
1. Change the base URL
2. Update your API key
3. Update model names
### Python
**Before (OpenAI)**:
```python
from openai import OpenAI
client = OpenAI(
api_key=os.environ.get("OPENAI_API_KEY")
)
response = client.chat.completions.create(
model="gpt-4",
messages=[{"role": "user", "content": "Hello"}]
)
```
**After (xAI)**:
```python
from openai import OpenAI
client = OpenAI(
api_key=os.environ.get("XAI_API_KEY"),
base_url="https://api.x.ai/v1" # Add this line
)
response = client.chat.completions.create(
model="grok-4", # Change model name
messages=[{"role": "user", "content": "Hello"}]
)
```
### Node.js/TypeScript
**Before (OpenAI)**:
```typescript
import OpenAI from 'openai';
const client = new OpenAI({
apiKey: process.env.OPENAI_API_KEY
});
```
**After (xAI)**:
```typescript
import OpenAI from 'openai';
const client = new OpenAI({
apiKey: process.env.XAI_API_KEY,
baseURL: 'https://api.x.ai/v1' // Add this line
});
```
### Model Mapping
| OpenAI Model | xAI Equivalent |
|--------------|----------------|
| gpt-4 | grok-4 |
| gpt-4-turbo | grok-4-fast |
| gpt-3.5-turbo | grok-4-1-fast-non-reasoning |
| dall-e-3 | grok-imagine-image |
## Migrating from Anthropic
### SDK Deprecation Notice
**Important**: The Anthropic SDK compatibility is fully deprecated. Migrate to the Responses API, OpenAI SDK, or gRPC.
### Using OpenAI SDK Instead
```python
from openai import OpenAI
client = OpenAI(
api_key=os.environ.get("XAI_API_KEY"),
base_url="https://api.x.ai/v1"
)
response = client.chat.completions.create(
model="grok-4",
messages=[
{"role": "system", "content": "You are a helpful assistant."},
{"role": "user", "content": "Hello"}
]
)
```
### Model Mapping
| Anthropic Model | xAI Equivalent |
|-----------------|----------------|
| claude-3-opus | grok-4 |
| claude-3-sonnet | grok-4-fast |
| claude-3-haiku | grok-4-1-fast-non-reasoning |
## Configuration Changes
### Base URL
```
https://api.x.ai/v1
```
### API Key
Get your API key from the xAI Console.
```bash
export XAI_API_KEY="your-xai-api-key"
```
## Feature Differences
### Parameters Not Supported by Reasoning Models
Grok 4 does not support:
- `presencePenalty`
- `frequencyPenalty`
- `stop`
- `reasoning_effort`
Remove these parameters when migrating to Grok 4.
### New Features Available
After migrating, you gain access to:
- Built-in web search tools
- X (Twitter) search tools
- Code execution
- Collections search
- Voice API
- 2M token context (Grok 4)
## Testing Migration
### Parallel Testing
```python
def compare_providers(prompt):
# OpenAI response
openai_client = OpenAI(api_key=os.environ.get("OPENAI_API_KEY"))
openai_response = openai_client.chat.completions.create(
model="gpt-4",
messages=[{"role": "user", "content": prompt}]
)
# xAI response
xai_client = OpenAI(
api_key=os.environ.get("XAI_API_KEY"),
base_url="https://api.x.ai/v1"
)
xai_response = xai_client.chat.completions.create(
model="grok-4",
messages=[{"role": "user", "content": prompt}]
)
return {
"openai": openai_response.choices[0].message.content,
"xai": xai_response.choices[0].message.content
}
```
## Common Migration Issues
### 1. Parameter Errors
```
Error: presencePenalty is not supported
```
**Solution**: Remove unsupported parameters for reasoning models.
### 2. Model Not Found
```
Error: Model 'gpt-4' not found
```
**Solution**: Use xAI model names (grok-4, grok-4-fast, etc.)
### 3. Authentication Errors
```
Error: Invalid API key
```
**Solution**: Ensure you're using an xAI API key, not OpenAI/Anthropic.
## LiteLLM Migration
If using LiteLLM, update your configuration:
```python
from litellm import completion
response = completion(
model="xai/grok-4", # Prefix with xai/
messages=[{"role": "user", "content": "Hello"}],
api_key="your-xai-api-key"
)
```
## Best Practices
1. **Test thoroughly**: Validate outputs match expectations
2. **Update model names**: Use appropriate Grok models
3. **Remove unsupported params**: Clean up incompatible parameters
4. **Use native SDK**: Consider xAI SDK for full feature access
5. **Gradual migration**: Migrate incrementally, not all at once
